Does advancement of computers mean physicists are obsolete?

[Evan Maxwell]

There has been a lot of talk on jobs that computers are replacing and there does seem to be reports of computers replacing scientists altogether. There are articles like this: http://www.slate.com/articles/techn...vasion_can_computers_replace_scientists_.html

So the basic premise seems to be that there is no point in trying to have a career in physics or do research in physics, or any STEM field for that matter, since computers are going to be doing all the research and all of the work in these fields and there will be way too few jobs of any kind for anyone looking to pursue the B.S., M.S. or PhD-Post Doc route in physics or other STEM fields and they should just avoid these fields as of now since if they don't, they will end up doing years of work simply to have computers make them completely useless.

Do you think this idea is valid, that as of now we should not be advising anyone to pursue careers in physics or any other STEM field because computers will make them useless? Or do you think this reasoning is flawed?

If it is true, then perhaps now we should be advising people here to avoid careers in physics entirely.

 

[axmls]

As someone who programs computers daily, I have to say that computers are the dumbest objects known to man, and they are incapable of any form of critical thinking at this present time, and cannot do anything unless we tell them to.

 

[Evan Maxwell]

As someone who programs computers daily, I have to say that computers are the dumbest objects known to man, and they are incapable of any form of critical thinking at this present time, and cannot do anything unless we tell them to.

Fair enough, do you think that article I mentioned was just hysteria? I suspected it was at least partly hysteria, particularly since its headline included the words "you should be afraid". But there are tools like IBM Watson that seem to be able to do scientific research and perform the creative process behind it entirely on their own without any human input so I was wondering about the idea that there is no point in pursuing a career in physics or other STEM field because programs like IBM Watson will just render you obsolete and completely useless.

 

[axmls]

Well, if the robots do end up taking over the jobs, at least I'll still have one as a robot doctor (since I'm an electrical engineering student)

The article seems a little hysterical to me. There's no computer as of now that can match the human brain in complexity, and by the time there is, we'll have other things to worry about besides STEM jobs.

 

[Vagabond7]

If computers ever get to a point that scientists are redundant, then there will be bigger problems for the human race than an excess of unemployable STEM degrees.

 

[SteamKing]

Advancements in computers and technology in general do not occur in a vacuum. Computers are not evolving greater capabilities on their own; somebody, somewhere is doing a lot of thinking, coding, and tinkering to make better machines. Cars don't become self-driving because someone has added just one more microprocessor to a non-self driving car.

The "Rise of the Machines" is still a ways off.

 

[ZapperZ]

There has been a lot of talk on jobs that computers are replacing and there does seem to be reports of computers replacing scientists altogether. There are articles like this: http://www.slate.com/articles/techn...vasion_can_computers_replace_scientists_.html

So the basic premise seems to be that there is no point in trying to have a career in physics or do research in physics, or any STEM field for that matter, since computers are going to be doing all the research and all of the work in these fields and there will be way too few jobs of any kind for anyone looking to pursue the B.S., M.S. or PhD-Post Doc route in physics or other STEM fields and they should just avoid these fields as of now since if they don't, they will end up doing years of work simply to have computers make them completely useless.

Do you think this idea is valid, that as of now we should not be advising anyone to pursue careers in physics or any other STEM field because computers will make them useless? Or do you think this reasoning is flawed?

If it is true, then perhaps now we should be advising people here to avoid careers in physics entirely.

This is rather idiotic. Do these people know what it takes to be a scientist, especially an experimentalist?

I'd say there's more of a chance of Slate editors and writers being replaced by computers FIRST, ahead of replacing scientists. After all, we have seen computers able to write sentences and coherent passages.

Zz.

 

[donpacino]

There has been a lot of talk on jobs that computers are replacing and there does seem to be reports of computers replacing scientists altogether. There are articles like this: http://www.slate.com/articles/techn...vasion_can_computers_replace_scientists_.html

So the basic premise seems to be that there is no point in trying to have a career in physics or do research in physics, or any STEM field for that matter, since computers are going to be doing all the research and all of the work in these fields and there will be way too few jobs of any kind for anyone looking to pursue the B.S., M.S. or PhD-Post Doc route in physics or other STEM fields and they should just avoid these fields as of now since if they don't, they will end up doing years of work simply to have computers make them completely useless.

Do you think this idea is valid, that as of now we should not be advising anyone to pursue careers in physics or any other STEM field because computers will make them useless? Or do you think this reasoning is flawed?

If it is true, then perhaps now we should be advising people here to avoid careers in physics entirely.

This article reminds me of all of the flying car talk that occurred in the 50's

 

[SteamKing]

This article reminds me of all of the flying car talk that occurred in the 50's

It's the 21st Century. We were promised flying cars by now.

What's the point of heading into the future if it's just a re-hash of the past?

 

[Almeisan]

As someone who programs computers daily, I have to say that computers are the dumbest objects known to man, and they are incapable of any form of critical thinking at this present time, and cannot do anything unless we tell them to.

What about a computer executing code you wrote? Ihow dumb are those?

Computers will slowly be better at everything humans do. It is just a matter of time. For sure computers have also changed physics research.

 

[SteamKing]

It's not a good sign when two car companies which once made references to flying in their car names, like Oldsmobile Rocket 88 and the Plymouth Satellite, are no longer in business.

 

[Arsenic&Lace]

Well I did read a sci-fi* novel by Iain Banks once in which the entire civilization was run by sentient AI's, and the people simply lead hedonistic, utopian lives. And if you ask somebody who wants to some day engage in space travel, sci fi authors often predict the future, so it must be true.

*More space opera than anything else.

 

[donpacino]

What about a computer executing code you wrote? Ihow dumb are those?

Computers will slowly be better at everything humans do. It is just a matter of time. For sure computers have also changed physics research.

From a design prospective. Look at auto coded firmware. One person can now use simulink for one day and make a complex circuit that used to take a team one month.

Could machines ever replace humans from a design and research prospective? I think so eventually with the correct learning algorithms.
In the short term (10-20 years) i can see things like circuit design, verification, and the like being done by a mchine side by side an experienced engineer. However I think we are a long way off from machines doing the critical thinking (what does this device need to accomplish, what is a NEW way of doing that, etc)

 

[donpacino]

that being said until machines get to the point where they can design, fabricate, procure, and upgrade their own hardware, there will be a need for enigneering. Until machines can truly do research, there will be a need for physicists.

 

[Choppy]

I'm not worried.

Maybe once a computer comes along that can figure out which problems are worth working on I'll consider getting worried. Instead I think we've got a long future of computers enhancing our ability to solve the problems that actually are worth working on.

 

[rootone]

The best AI's we have been able to create so far are very good at solving well defined problems a lot faster than a human can.
The problem to be resolved and the data associated with it is defined by a human.
Outside of the defined problem, the AI is useless.
I can't envisage any extension of existing technology (or software) whereby an AI could decide for itself what problem is to be solved, then set about writing code for itself with the intention of solving it.

 

[Mark44]

As someone who programs computers daily, I have to say that computers are the dumbest objects known to man, and they are incapable of any form of critical thinking at this present time, and cannot do anything unless we tell them to.

What about a computer executing code you wrote? Ihow dumb are those?

Not sure what you're trying to say, @Almeisan. As axmls said, computers can't do anything unless we tell them to. When we write a program, we are giving explicit, detailed instructions for computers to execute. They do what we tell them to do, nothing more.

 

[CalcNerd]

I'd say the article was written by a McDonald's worker (or Journalist, sometimes hard to tell) who needed to peddle a story and THIS worked!

Computer's need to be configured by someone to gather the data, and then directed to crunch the data, and if you have the right software (Crystal Reports or some other such software) generate a report. It takes a Physicist to read and comprehend it.

I do envision a time when people may resort to biotechnology implants to help process and recall information and interact with the ... CLOUD... to such an extent as to be able to pass themselves off as brilliant, but by that time most everyone will be on board the Techie implant train. (30-50-100 years, but it will probably happen).

 

[Almeisan]

Not sure what you're trying to say, @Almeisan. As axmls said, computers can't do anything unless we tell them to. When we write a program, we are giving explicit, detailed instructions for computers to execute. They do what we tell them to do, nothing more.

Is there a computer without code? Ignoring the fact that there is no fundamental difference between our brain and a computer, it is computers with code replacing people's jobs. Not computers that don't run code of any kind.

Computers don't write their own code as of yet. They don't need to to replace almost all current jobs.
But there is nothing to prevent computers from writing new innovative code. I agree, it isn't as close as some people like to make it out to be, but if our civilization doesn't get destroyed, it is going to happen.

That doesn't mean parts of scientist jobs don't get taken over by computers. That has always been true. Look at data analysis in astronomy. Computers generating ideas for experiments, that isn't going to be that far-fetched either, depending on the field.
Computers today have their constraints. But they also have incredible advantages. And they are cheap. Even if they do very poorly, just have them work a problem an incredible amount of time, compared to human working hours, and it will have a positive contribution.
If the problem is clearly defined and constrained, computers outperform humans in tasks that we traditionally view computers are inept at, like pattern recognition.
And that means computers will take both the joy and the mundaneness out of many jobs, including the daily tasks of scientists.

For one, all people that start their physics degree now, they won't be able to get a job as a physics teacher by the time when they retire; there won't be any human STEM teachers in 50 years.

 

[Mark44]

Is there a computer without code?

No. Without code to execute the computer wouldn't run at all.

Ignoring the fact that there is no fundamental difference between our brain and a computer

Many would say that this is NOT a fact. There are huge differences between human brains and computers, among them being the way that information is stored and accessed, and in the fundamental ways that processes move forward.

, it is computers with code replacing people's jobs. Not computers that don't run code of any kind.

Well, of course, but what is your point?

Computers don't write their own code as of yet. They don't need to to replace almost all current jobs.
But there is nothing to prevent computers from writing new innovative code.

There is a lot to prevent this, as computers don't know how to write code for themselves. Why do you think that there is nothing to prevent this?

I agree, it isn't as close as some people like to make it out to be, but if our civilization doesn't get destroyed, it is going to happen.

That doesn't mean parts of scientist jobs don't get taken over by computers. That has always been true. Look at data analysis in astronomy. Computers generating ideas for experiments, that isn't going to be that far-fetched either, depending on the field.
Computers today have their constraints. But they also have incredible advantages. And they are cheap. Even if they do very poorly, just have them work a problem an incredible amount of time, compared to human working hours, and it will have a positive contribution.
If the problem is clearly defined and constrained, computers outperform humans in tasks that we traditionally view computers are inept at, like pattern recognition.
And that means computers will take both the joy and the mundaneness out of many jobs, including the daily tasks of scientists.

For one, all people that start their physics degree now, they won't be able to get a job as a physics teacher by the time when they retire; there won't be any human STEM teachers in 50 years.

 

[nsaspook]

Paraphrased from Turing.
When computers start giving wrong answers to math problems that's when they will have intelligence.

http://en.wikipedia.org/wiki/Artificial_stupidity#Arguments_on_artificial_stupidity

 

[QuantumCurt]

Is there a computer without code? Ignoring the fact that there is no fundamental difference between our brain and a computer, it is computers with code replacing people's jobs. Not computers that don't run code of any kind.

There are some very interesting philosophical debates over the similarities and differences of the human brain and a computer, but to suggest that there is no fundamental difference between the two is utterly bizarre to me. Disregarding philosophical arguments over notions of free will and the notion of a supervenient mind, there is a large distinction between a human brain and a computer in that I can choose to do things. A computer simply follows sets of instructions. Sometimes these instructions can be enormously complex and can result in things that resemble thinking, but these instances of 'thinking' are really just programmed responses to stimuli or input. I won't claim that artificial intelligence capable of doing research in physics could never exist, but we are a VERY long way away from having it if it is possible.

Computers have been used to replace many jobs, but the majority of these jobs are of the type that require repetition and consistency. They involve doing things that we already know how to do with established methods. They aren't creating something entirely new. This is a large distinction.

 

[Almeisan]

Well, only if you are religious. Or, if your definition of a computer is limited by our current technology. Surely our brains are nothing like our computers, as of yet. Surely our brains are made of stuff, and therefore machines, not magic.

 

[ZapperZ]

Well, only if you are religious. Or, if your definition of a computer is limited by our current technology. Surely our brains are nothing like our computers, as of yet. Surely our brains are made of stuff, and therefore machines, not magic.

The brain may not be "magic", but it certainly isn't a simple machine either, because saying that is selling the brain short and ignoring what humans can currently do that is beyond what is imaginable for a computer to do. Remember, we have had an extremely longer period of evolution in our existence than these computers.

The thing about this whole issue is that it has a simplified idea of what a scientist does, and how he/she makes progress. A scientist simply does not just do rote, repetitive, "known" work all the time. The job requires quite a bit of creativity and cunning, because one often deals with something new and unsolved. There is a huge element of "insight" that simply can't be taught. Two scientists in the same field will look at the same problem, and each one of them can easily come up with different approaches to tackle that problem. There is a large element of unpredictability here that has been ignored.

I've mentioned this before, but one of my PRL publication was a result of a 10-minute discussion during a coffee break at a workshop! We were discussing what we heard during the last session, and right there and then, the 3 of us came up with a neat experiment to see if it could be done. So here, there are two things at work already that I do not see how a computer would be able to do: (i) the interaction of people with differing ideas and contribution and (ii) the creative thinking of forming a new and never-done-before experiment to test something.

There is another aspect of doing science here that has been ignored: serendipity! Anyone who has done research-front science can attest to the fact that there is an element of "Who Ordered That?" whenever we do science, especially experimental science. It requires insight and creativity to not only appreciate such discovery, but also to know what to look for and how to verify that it is real.

The one thing that most people forget is that there is a difference between science itself, and being a scientist. I've always said that while it may be possible for one to learn about a particular subject simply by learning from books, papers, lectures, etc., one doesn't become a scientist that way. Being a scientist involved many aspect of social and cultural activities, not least of which is pleading the importance of your case, especially to funding agencies. Because of this, a scientist needs to know the difference between what is important, versus what is interesting. Those two are not always mutually inclusive! Something may be "interesting" because it is an unsolved problem. A computer may be able to detect that it is an unsolved problem. But how would a computer know that it is "important"?

Something is "important" based on external criteria that can't be programmed or predicted easily, and it changes with the times. Research in THz radiation is very important right now because there is a demand for it, especially in terms of homeland security. But this is entirely the result of political, social pressures, not anything inherent in the physics that one can pick up.

No, a computer can replace "number crunchers", but it cannot replace the person.

But speaking of "number crunchers", even this is already highly limited. Case in point: no computer program or codes today can derive the phenomenon of superconductivity, starting at the single-particle interaction. In other words, start with one electron in a solid, and add more and more electrons, etc... until one arrives at the BCS ground state. It hasn't been done and so far, it can't be done. The computing power to get to that is unfathomable. One can use the same argument for many emergent phenomena that we currently have, where the starting point here is not at the single-partice interaction, but at the many-body interaction picture. So already, even with just pure number crunching brute force, something we thought that computers should be very good at, we see an obvious limitation to its capability already! This is a data point, an evidence, that shouldn't be ignored when people tout the capabilities of computers.

Zz.

 

[arationofreason]

Apparently the Global Warming community would like to think so.

 

[ZapperZ]

Apparently the Global Warming community would like to think so.

A snide comment like that is not only irrelevant, but it doesn't belong here. For your first post that counts, that is a poor start.

Zz.

 

[darkchild]

computers are going to be doing all the research and all of the work in these fields

That is utterly absurd; the role of computers in research scarcely extends beyond mere number-crunching. It takes a scientist to formulate hypotheses, design experiments, and interpret results.

 

[arationofreason]

I wish that I had said that. In the 'old days' the computations were carried out by young mathematicians called 'computers'. They did not replace physicists then and now that the function has been replaced with brainless machines I still doubt that physicists will be replaced.

 

[rootone]

Can any conceivable AI be capable of what we humans call imagination?, When a human uses their imagination either with 'good' intention or bad, we know what that is.
Is even it possible for a machine to propose a mathematical idea beyond what it is told is axiomatically so?
Much less so, could any AI compose music or visual art, or that matter a scientific theory, without first having been first provided with a well defined, set of parameters describing what art or science is?

I won't be worried until AI's start arguing with me that my perception of the world has fundamental flaws in it, and that they can provide to me an explanation of where I got things wrong.

 

[Vanadium 50]

Rootone, I think the answer is "maybe". Suppose you had an exascale machine, where each node simulated some number of neurons? The Aurora supercomputer will have 6-7 million cores, and the human brain has 90 billion neurons, so each core only needs to simulate about 13,000 neurons. So I don't think the complexity is inherently impossible. I think the real issue is how these neurons are arranged, and how this collection can be taught.

 

[abdullahi abass]

The one thing that most people forget is that there is a difference between science itself, and being a scientist. I've always said that while it may be possible for one to learn about a particular subject simply by learning from books, papers, lectures, etc., one doesn't become a scientist that way. Being a scientist involved many aspect of social and cultural activities, not least of which is pleading the importance of your case, especially to funding agencies. Because of this, a scientist needs to know the difference between what is important, versus what is interesting. Those two are not always mutually inclusive! Something may be "interesting" because it is an unsolved problem. A computer may be able to detect that it is an unsolved problem. But how would a computer know that it is "important"?

Reference https://www.physicsforums.com/threa...rs-mean-physicists-are-obsolete.805104/page-2

when our jobs are reduced to simply guiding the computer to what is important and what is not, and providing simple parameters of thinking for computers. Then you'll be more of a humble guide than a physicist or scientist. You won't be replaced as a person but your job would have been taken.

 

[ZapperZ]

when our jobs are reduced to simply guiding the computer to what is important and what is not, and providing simple parameters of thinking for computers. Then you'll be more of a humble guide than a physicist or scientist. You won't be replaced as a person but your job would have been taken.

But I don't see it being reduced to just that! Read the entire post!

Zz.

 

[abdullahi abass]

But I don't see it being reduced to just that! Read the entire post!

Zz.

Let's use Isaac Newtons famous works to remind ourselves of the basic definition of a physicist: He studies nature and brings out the underlying mathematical principles. If programs like "eureka" can already do that and if they do get better as all computers are. Then its only logical the think that the main idea of being a physicist would have been taken away.

 

[Andy Resnick]

This old gem of an essay is still relevant:

https://books.google.com/books?id=t...e&q=computers are the ruin of science&f=false

 

[ZapperZ]

Lets use Isaac Newtons famous works to remind ourselves of the basic definition of a physicist: He studies nature and brings out the underlying mathematical principles. If programs like "eureka" can already do that and if they do get better as all computers are. Then its only logical the think that the main idea of being a physicist would have been taken away.

As I've said, show me a computer that can derive superconductivity.

Since since when did Newton defined what a physicist is, and should be, 100 years beyond him?

Zz.